Communications of the ACM
Beware the Lure of the Horizontal
One of the more useful growth and diversification strategies for software companies is to segment markets horizontally or vertically. When we use the word "horizontal" in reference to a software market, that usually refers to a segment or product type that potentially appeals to most or all computer users, regardless of their industry or functional specialization. Horizontal products include operating systems, browsers, other systems software such as security programs, as well as basic applications like Microsoft Office, Oracle databases, or Intuit's Quicken personal finance program.
By "vertical," we usually mean a market segment that targets a much more specific set of users. The vertical domain may be defined by an industry (for example, software for the health care industry, which is likely to be a suite of otherwise-horizontal database, procurement, accounting, scheduling, and other programs tailored to the requirements of hospitals or doctors' office managers). It could be a technical specialty (such as computer-aided design programs or embedded controller software) or a technical specialty for a particular industry (such as computer-aided design programs for semiconductor design engineers, or apparel designers, or automobile designers). It could also be a platform-specific market (for example, an application that runs only on a particular operating system and computer hardware combination).
Many software companies target horizontal segments because these are more likely to be mass markets and generate much larger sales than specialized products for vertical markets. The number of potential customers for a horizontal product is roughly equal to the number of individuals or organizations with computers. But while horizontal products can make some firms billions of dollars, managers need to beware of numerous downsides.
One negative is that horizontal markets can attract lots of competitors. Too many players can lead to stiff price competition to the point where few companies make any money.
Developers of horizontal products like operating systems and communications software quickly discover that truly homogeneous sets of users are rare.
A second downside of horizontal market strategies is the cost, and associated risks. It is one thing to imagine selling a product to every individual user or corporate customer. But it is another thing to invest in designing, developing, selling, and supporting a product that accommodates so many different types of users in an almost infinite set of user scenarios.
Developers of horizontal products like operating systems and communications software quickly discover that truly homogeneous sets of users are rare. They often end up adding lots of features to their products to attract the broadest number of users. But the products can turn out to be expensive to produce and update.
A third risk is that the horizontal product turns out to have features that make it a "lowest-common-denominator" product for some theoretical ideal of an average user but not really the most common user. The result is the software product does not sell in volume because it is too weak in functionality compared to more specialized products.
A fourth issue is that the lure of the horizontal can lead companies to a strategy whereby they bundle many separate products together to create a broad suite or solution that they believe will appeal to a large number of users. This strategy can sometimes work brilliantlyas in the case of Microsoft with its Office product. It seems that almost every enterprise software company, ranging from Oracle to SAP, offers a bundle of products that provide a "solution" to something, like putting your business online.
The problem here is that a company may introduce weak products to create this horizontal-looking solution, but end up using the stronger products to subsidize the weaker ones. Every producteven if few people buy itrequires money and people for software design, engineering, testing, documentation, and product support as well as sales and marketing. This portfolio variance occurs in most companiesfew firms with broad product lines have offerings that are equally strong. But when it occurs too frequently in the same firm, it suggests an organization that has been unable to resist the lure of the horizontal and that has expanded unrealistically and unwisely.
Even great companies can vastly overestimate their ability to sell horizontal products. Microsoft is the most successful example of a software company that has focused exclusively on horizontal markets. In its last fiscal year, it had revenues of more than $28 billion and an operating profit rate of 42%. The company sold some 60 separate software products (though some are sold only in bundles) in four major categories (desktop applications, desktop platforms, enterprise software, and consumer software, including mobile and embedded systems). These are all potentially large horizontal software markets. Microsoft is always on the lookout for new horizontal markets as well. But very few of Microsoft's products have high enough prices or enough users to make them profitable.
Based on new financial reporting data for fiscal 2003, it appears that only three of Microsoft's product linesWindows desktop, Windows servers, and Officeaccount for more than 80% of the company's revenues and all of its profits. The other product areas were losing money in 2003. Although some of its other offerings will undoubtedly become larger sources of revenues and profits in the future, Microsoft would be about 50% more profitable if it concentrated only on these three product areas.
Rather than continuing to segment markets horizontally, another strategy is for software firms to do the oppositeblur the distinctions across horizontal boundaries and create new generations or types of products with broad appeal. One of the reasons why Windows continues to sell in the upgrade market is that Microsoft has been successfully adding different systems, applications, and networking or communications features to the core product. Unfortunately, Microsoft has damaged the businesses of many former horizontal players making products like screen savers, file management utilities, data-compression programs, browsers, and other networking and communications products. But the strategy has worked for Microsoft, at least when it has not violated antitrust law. In addition to Microsoft, other software companies like Netscape, Sun Microsystems, IBM, and BEA Systems have built enormous horizontal businesses by creating new categories of products such as Web application servers that sit in between systems, networking, and applications software.
Another alternative for an enterprise software company is to combine horizontal with vertical market segmentation. In other words, develop general-purpose products that could be on every desktop (or enterprise-class server), and then get an extra boost in the marketplace by creating specialized versions of these products tailored for particular industries or types of users.
Many companies in the enterprise resource planning realm routinely exploit horizontal and vertical market segmentation. One of the best examples is SAP, Europe's largest software company and one of the top 10 software companies in the world by revenues. The company in 2002 had revenues of approximately $6.88 billion and an operating profit rate of 23%. It employed over 28,000 people and operated in 50 countries. Founded in Germany in 1972 by five former IBM systems engineers, SAP initially developed a series of back-office software packages designed to handle different horizontal enterprise functions such as finance, personnel, supplier management, procurement, and sales management. Because firms in each industry tend to have slightly different ways of treating these functions, most enterprise software companies have to offer extensive customization services or rely on partners to tailor their products. SAP, however, decided to preempt some of this customization by creating specialized versions of its horizontal products tailored for various vertical industry segments. The company still has enormous revenues from services and maintenance upgrade sales (about 70% of 2002 revenues), but many of these sales come from training, installation, and integration work, rather than customizing code for individual users. For the most part, SAP encourages users to adopt procedures and data definitions built into its products, rather than adapt the software to user procedures and data.
The horizontal-vertical combination also makes sense because pure horizontal products can require enormous investment and skill to master. The lure is so great that these kinds of markets tend to attract software startups that don't have the resources of a Microsoft or an SAP and shouldn't be playing the horizontal game. Many companies have tried, for example, to be "the next Microsoft" in the wireless space or the handheld computing space by producing operating systems, development tools, and a myriad of applications for handheld cell phones and PDA devices. Very few companies will succeed.
On the other hand, startups often can develop versions of their products that are just right for a particular vertical market, such as a mobile operating system optimized for real-time applications or a Web-based accounting package optimized for medium-sized firms in the retailing business. It must be easier to design and sell a product exactly right for a particular set of customers rather than for all computer users regardless of industry or specific customer scenarios. Moreover, if a company can master one vertical market, then perhaps it can master another one, and another one (like SAP did). By growing in increments, a company might eventually turn a series of vertical solutions into a horizontal product with broader appeal. But a new software venture needs to start somewhere and prove itself before tackling the world. A narrow but attractive vertical market seems like a good place to start.
©2003 ACM 0002-0782/03/0700 $5.00
Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, to republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee.
The Digital Library is published by the Association for Computing Machinery. Copyright © 2003 ACM, Inc.